2

I am reading Aluffi's "Algebra: Notes from the Underground" book and one of the exercises in the books asks the following:

Let $R$ be a commutative ring, and let $r_1, \ldots , r_n$ be elements of $R$.
Prove that $R[x_1,\ldots,x_n] / (x-r_1,\ldots,x-r_n) \cong R$.

I have tried to find an explicit isomorphism between the two sets in a similar fashion as to how one shows that $R[x] / (x-r) \cong R$, namely by using the function $e_r : R[x] \rightarrow R, f(x) \mapsto f(r)$, the First Isomorphism Theorem and the fact that $\ker(e_r)=(x-r)$ but I haven't been able to find a multivariable equivalent to $e_r$.

I have also tried the following: I know that $R[x_1,\ldots,x_n] \cong R[x_1,\ldots,x_{n-1}][x_n]$ and also that $R[x_1,\ldots,x_{n-1}][x_n] / (x-r_n) \cong R[x_1,\ldots,x_{n-1}]$ so $R[x_1,\ldots,x_n] / (x-r_n) \cong R[x_1,\ldots,x_{n-1}]$. Similarly, $R[x_1,\ldots,x_{n-1}] \cong R[x_1,\ldots,x_{n-2}][x_{n-1}]$ and $R[x_1,\ldots,x_{n-2}][x_{n-1}] /(x-r_{n-1}) \cong R[x_1,\ldots,x_{n-2}]$ so $$\bigl(R[x_1,\ldots,x_n]/ (x-r_n)\bigr)/(x-r_{n-1}) \cong R[x_1,\ldots,x_{n-2}] $$ and here I get stuck again, as I don't know how to make sense of the left-hand side of the last isomorphism.

So, I would appreciate if someone would tell me how to reason about this problem. Thanks.


Note that this book adopts a rings-first approach and the tools available to the reader at this point reach the First and Third Isomorphism Theorem level.

lorenzo
  • 4,298

2 Answers2

3

Your approach is correct. For ideals $I,J \subseteq R$ of a commutative ring $R$ it holds $R/(I + J) \cong (R/I)/((I+J)/I)$. For example for $n = 2$: $$R[x_1,x_2]/(x_1-r_1,x_2-r_2) = R[x_1,x_2]/((x_1-r_1)+(x_2-r_2))\\ \cong (R[x_1,x_2]/(x_2-r_2))/((x_1-r_1)/((x_1-r_1)+(x_2-r_2))) \\ \cong R[x_1]/(x_1-r_1) \cong R.$$ The general case follows similarly by induction.

An explicit isomorphism is given by $R[x_1,\dots,x_n]/(x_1-r_1,\dots,x_n-r_n) \to R, \ f \mapsto f(r_1,\dots,r_n)$.

psl2Z
  • 5,772
  • 1
    Please don't answer duplicate questions as this is against the site guidelines. Instead help to find and link the duplicate. Basic questions like these have been asked 10 times here over the years. – Martin Brandenburg Nov 14 '24 at 00:23
3

Here is a different approach, that avoids induction. There is a natural surjective homomorphism $\varphi:R[x_1,...,x_n]\to R$ given by $f\to f(r)$ where $r=(r_1,...,r_n)$. Let $I=(x_1-r_1,...,x_n-r_n)$. We just have to show that $I=\ker(\varphi)$. Clearly, $I$ is contained in the kernel. As for the reverse inclusion, there is a nice trick. Let $f\in\ker(\varphi)$, i.e $f(r)=0$. Write explicitly $f=\sum a_{i_1,...,i_n}x_1^{i_1}...x_n^{i_n}$. Now, by the definition of $I$, in the quotient ring $R[x_1,...,x_n]/I$ we have $x_i+I=r_i+I$ for each $i$. Hence:

$f+I=\sum a_{i_1,...,i_n}x_1^{i_1}...x_n^{i_n}+I=\sum a_{i_1,...,i_n}r_1^{i_1}...r_n^{i_n}+I=f(r)+I=0+I=I$

Which means $f\in I$, as required.

Mark
  • 43,582
  • 1
    Please don't answer duplicate questions as this is against the site guidelines. Instead help to find and link the duplicate. Basic questions like these have been asked 10 times here over the years. – Martin Brandenburg Nov 14 '24 at 00:24